A preliminary study on cause‑of‑death discrimination and the pathological stage identification in acute ischemia heart disease (AIHD) based on plasma lipidomic technique and machine learning algorithms.
Journal:
International journal of legal medicine
Published Date:
Jun 6, 2025
Abstract
The sudden death discrimination of acute ischemia heart disease (AIHD) and the determination of the AIHD pathological stage are the difficulties in forensic medicine. More potential biomarkers with high sensitivity and specificity still need to be identified to diagnose AIHD. Current studies have linked concentration variation in lipid characteristics after death to the identification of causes of death, providing a potential strategy for diagnosing AIHD. In this study, we used ultra-high performance liquid chromatography-mass spectrometry (UHPLC-MS/ MS) to systematically analyze the non-targeted lipid metabolism profile of the corpse blood of AIHD and non-cardiac disease death cases. A total of 665 lipid metabolites were detected. The study rigorously analyzed the performance of 8 cutting-edge machine learning algorithms in accurately identifying AIHD. We identified 18 lipid metabolites for AIHD discrimination and 47 for early myocardial ischemia (EMI) and acute myocardial infarction (AMI) identification according to the feature importance. We developed an e-Xtreme gradient boosting (XGB) optimized classification model (AUC = 0.830, Accuracy = 0.781) and a logistic regression (LR) optimized model (AUC = 0.990, Accuracy = 0.964). Our results demonstrate the potential application of plasma lipidomic technique combined with machine learning in diagnosing the cause of death and determining the pathological stage of AIHD.
